A Linear Data-Driven System Identification Methodology for an Active/Passive Solar Thermal Storage System and Application to a Solar House

摘要

This thesis presents a methodology developed to identify a parametric model of a thermally-activated building system (TABS) using a system identification (SI) tool. The model was identified with collected data from an energy efficient solar single-family residential building, the Ecoterra™ house located in Eastman, Quebec. The TABS is a ventilated concrete slab (VCS) serving as an energy storage medium for active and passive solar gains in the basement. The system uses the structural mass of the house to store active solar gains collected by the building-integrated photovoltaic/thermal (BIPV/T) roof and passive solar gains entering the living space through the energy efficient windows.udA data-driven system identification approach is used. Identifying a linear model and obtaining a low-order polynomial model were the main identification criteria. The thesis addresses the issues of the monitoring sensor accuracy on the model parameters, how physical knowledge of the VCS dynamic system can be considered during its validation and how the identification assumptions guide the future use of the model.udThis thesis also demonstrates that the identified linear polynomial model is an efficient tool to carry out redesign studies and possible control studies. Improved BIPV/T roof designs are compared based on increased solar energy utilization potential and potential increase of collected thermal energy stored into the VCS. The effect of modifying the BIPV/T roof angle and including a glazing section are analyzed and discussed, demonstrating the use of the identified transfer function model of the VCS as a more efficient and quicker alternative to whole building detailed simulations, particularly for comparing design and operation options on a relative basis.ud